B737 Throttle Quadrant - Automated Thrust Lever Movement

Autothrottle arming switch is a solenoid operated switch clearly identified on the main Instrument Panel (MCP).  The switch is linked to the IAS/MACH speed window (adjacent) and to two A/T disconnect buttons located either side of the throttle lever handles. cp flight pro mcp

In this final post dealing with the conversion of the 737-500 throttle quadrant. I will discuss the automation and movement of the throttle thrust levers and touch on some problems that occurred.  I will also briefly discuss the installation and use of potentiometers.  Part of this post will be repetitive as I briefly discussed automation in an earlier post.

Avoiding Confusion - Automation

To avoid confusion, automation refers only to the movement of the two throttle thrust levers in relation to the %N1 output.  These N1 limits and targets are provided by the Flight Management Computer (FMC) and normally are used by the Autopilot Flight Director System (AFDS) and the Auto Throttle (A/T) to maintain airspeed and/or thrust setting.  

Automation and Movement - Interface Cards

Automation is the use of CMD A or CMD B (autopilot) to control the %N1 outputs from the Autothrottle (logic), and motorization is the moving of the throttle levers in unison with %N1 output.  To achieve this seamlessly, two interface and one controller card are used.

Alpha Quadrant Cards (2):  Each  motor controller card has the automation logic programmed directly to the card (via propriety software).  One card controls Auto Pilot CMD A while the other card controls Autopilot CMD B.

Phidget Advanced Servo Card (2):  This card acts as an interface and bridge between the Alpha Quadrant cards and the flight simulator platform used.

The card does not provide movement for the throttle thrust levers; this is controlled by a Phidget Motor Controller card.

Leo Bodnar BUO 836 A Joystick Controller Card:  This card will register in Windows the movement of levers, buttons and switches on the TQ.  Calibration of this card is done first in Windows, then in Flight Simulator (FSX/P3D), FSUIPC or the avionics suite used; for example, ProSim737.

The interface cards are mounted forward of the MIP within the Throttle Interface Module (TIM) and are connected to the throttle unit by custom VGA straight-through cables and to the computer by a single USB cable.

Main Controller Cards

The controller card I have used is not a Phidget card but a specialist card often used in robotics (Alpha Quadrant card).  The software to program the card has been independently developed by a software engineer and does not utilize Phidgets.

The technology used in the controller card is very similar to that utilized by NASA to control their robotic landers used in the space industry.  The technology is also used to control robots used in the car industry and in other mass production streams.  One of the benefits of the card is that it utilizes a software chip (firmware) that can be easily upgraded, reprogrammed, or replaced.  

The Alpha Quadrant cards provide the logic to operate the throttle automation (the movement of the thrust levers) and act as a bridge between the two cards and the avionics suite.

Being able to program each card allows replication of real aircraft logic and systems.  Whenever possible, these systems and their logic have been faithfully reproduced.

oem throttle. toga switches clearly seen

CMD A/B Autopilot - Two Independent Systems

Most throttle units only use one motor controller card which controls either CMD A or CMD B; whichever autopilot you select is controlled by the same card.  

In the real aircraft to provide for redundancy, each auto pilot system is separate.  This redundancy has been duplicated by using two Alpha Quadrant controller cards, rather than a single card.  Each controller card has been independently programmed and wired to operate on a separate system.  Therefore, although only one CMD is operational at any one time, a completely separate second system is available if CMD A or B is selected on the MCP.

Synchronized or Independent Lever Motorization

Synchronization refers to whether the two throttle thrust levers, based upon separate engine %N1 outputs, move in unison with each other (together) or move independently.

In earlier Boeing aircraft, such as the 707, 727 and 737 classics, the levers were roughly synchronized; however, the Next Generation has as a computer-operated fuel control system which can minutely adjust the %N1 of each engine.  This advanced fuel management can be observed in a real aircraft whereby each throttle lever creeps forward or aft independent of the other lever.

Programming flight simulator to read separate %N1 outputs for each engine, and then extrapolating the data to allow two motors to move the throttle levers independently is possible; however, the outputs are often inaccurate (for varying reasons).  This inaccuracy can often be observed on reproduction throttle units that exhibit a gap between lever one and lever two when automating %N1 outputs.  

It was decided to maintain the older system and have both levers synchronized.  Although this is not replicating the Next Generation system, it does make calibration easier.  If in the future incremental thrust lever movement is required, it’s a matter of adding another 12 Volt motor to the front of the throttle bulkhead to power the second thrust lever.  

Autothrottle activation will advance both thrust levers in unison to the fmc calculated %N1 output

Be aware that although both thrust levers are synchronized, the throttle handles may still show a slight difference in position in relation to each other.  This is caused by the varying tension that needs to be maintained on the fan belt connecting the 12 Volt motor to the mechanical system beneath the thrust levers.

Another aspect to note is that the position of the thrust levers during automation is arbitrary and is a visual representation of the %N1 output; it may or may not reflect the exact position on the throttle arc that the thrust lever would be placed if moved manually (by hand with Autothrottle turned off).  

Although the throttle is automated, manual override (moving the thrust levers by hand) is possible at any time, provided the override is within the constraints of the aircraft logic (programmed into the Alpha Quadrant card), and that provided by the flight avionics (ProSim737).  

Power Requirements and Mechanics

To provide the power to move the throttle thrust levers, a 12 Volt motor previously used to power electric automobile windows, is mounted forward of the throttle bulkhead (see image at bottom of post).  Connected to the motor's pulley is a fan belt that connects to the main pulley located beneath the thrust levers.  To enable the thrust levers to move in unison, a slip clutch, which is part of the main pulley assembly, is used.  

Captain-side TO/GA button is clearly seen below lever handles.  The button at the end of the handle is the Autothrottle disconnect button

ProSim737 Limitations - TO/GA and Auto Throttle Override

Unfortunately, concerning automation the ProSim737 is deficient in two areas: TO/GA and A/T Override.

(A)  TO/GA

In the real aircraft, the flight crew advances the thrust levers to power 40%N1 (or to whatever the airline policy dictates), allows the engines to spool, and then pushes the TO/GA button/s.  Pressing TO/GA causes the throttle to go on-line and to be controlled by the AFDS logic.  The throttle levers then advance automatically to whatever %N1 the logic deems appropriate based on takeoff calculations.

As at the time of writing. If you're using ProSim737, this will NOT occur.  Rather, you will observe the thrust levers retard before they advance (assuming you have moved the thrust levers to %40 N1).  The reason for this is nothing to do with how the throttle is calibrated, FSUIPC or anything else.  ProSim737 software controls the %N1 outputs for the automation of the thrust levers and the developer of the software has not fine-tuned the calibration in the software to take into account real-world avionics logic.  This thread located on the ProSim737 forum provides additional information. 

I have not tested Sim Avionics, but have been told this issue doesn't occur in their avionics suite.

There are two workarounds:  Engage TO/GA from idle (hardly realistic) or push the thrust levers to around 80% N1, allow the engines to spool, then push TO/GA.  Anything less that around 80% N1 will cause the thrust levers to retard before advancing.

Latest ProSim737 release (V133)

The latest version of ProSim737 (V-133) has provided improvement to the above issue.  Throttles can now be advanced to ~60% N1 and TOGA engaged without the throttle levers retarding.  This is possible ONLY if you calibrate the throttle levers within ProSim and allow ProSim to control the throttle output logic.  if you calibrate within FSUIPC then the same issue will apply.

According to ProSim developers, this issue is probably related to the calibration of the ProSim servo output. When you press TO/GA, the current N1 is taken and calculated back to a throttle percentage. This throttle percentage, when combined with the servo calibration data from ProSim results in a servo output. The servo calibration at the moment only has 2 calibration points, which are 0 and 100%. This results in a linear behavior between the two points, while depending on the construction of the throttle, the relationship might be non-linear. This would require a multi point calibration which is hard to do at the moment, because a throttle does not have exact readouts of the current position, so it will be hard to calibrate a 50% point.

This may need improvement in the code to auto calibrate the throttle system.

It's hoped that future release of ProSim737 will rectify this issue.

(B) Autothrottle Manual Override

In the real aircraft, manual override is available to a flight crew and the thrust levers can be retarded with the Autothrottle engaged.  When the flight crew release pressure on the thrust levers the Autothrottle will take control again and return the thrust levers to the appropriate position on the throttle arc dependent upon the speed indicated in the speed window of the MCP.

ProSim737 will not temporarily disconnect (manual override) the Autothrottle.  

At the time of writing, there is no workaround to solve this.

Potentiometers

There are many types of potentiometers; the two types most commonly used in flight simulation are the linear and rotary types. Linear potentiometers are inexpensive, often have a +- percentage variance, are compact, have a minimal throw depending upon the size of the device, and are not immune to contaminates building up on their carbon track. 

The last point is worth mentioning, as it is wrongly assumed that a potentiometer will remain correctly calibrated for the life of the unit.  General wear and tear, dust, and other debris will accumulate on the potentiometer; any of which may cause calibration and accuracy problems.  Keeping the potentiometers free of dust is important.

Rotary potentiometers (which may have a string attached) are very accurate, are in a sealed case and have very minimum chance of contamination. They are also made too exacting standards, are larger in size, and are expensive.

To read further about potentiometers

QAMP secured to base of throttle unit.  Thumb screws are visible on each corner of the plate.  A possible add on modification to reduce the risk of dust contamination to the potentiometers is a plastic cover that fits over the plate (a lunch box)

Calibration of Potentiometers

The main method of calibrating the position of the thrust levers is by calibrating the potentiometer in Windows, then in FSX followed by fine tuning in FSUIPC (if needed).  

At the moment I am using linear potentiometers; therefore, at some stage cleaning or replacement may be required.   The 737 throttle quadrant is not cavernous and only certain sized potentiometers will fit into the unit; this combined with other parts and wiring means that the potentiometers are often inaccessible without removing other components.  

To allow speedier access to the potentiometers, a Quick Assess Mounting Plate was designed.

quick access mounting plate (QAMP). four linear potentiometers are mounted to the plate. Two grub screws secure the plate to the throttle chassis

Quick Access Mounting Plate (QAMP)

The potentiometers are mounted directly onto a custom-made aluminum plate that is attached to the inside of the throttle unit by solid thumb screws.   To access the plate, the side inspection cover of the throttle is removed (a few screws) followed by turning the thumb screws on the access plate.  This releases the plate.

A similar plate has been designed and constructed for use with the stand-by potentiometer that controls the flaps.  A more detailed picture of the QAMP can be seen here..

Below is a video showing the movement of the thrust levers with the autothrottle engaged recorded during a test flight. 

 
 

Teething Issues with the Throttle Conversion

It was envisaged that more problems would have surfaced than have occurred.  The major issues are outlined below:

(A) Trim Wheels

An early problem encountered was that the trim wheels when engaging generated considerable noise.  After checking through the system, it was discovered that the two-speed rotation of the trim wheels were causing the two nuts that hold each of the trim wheels in place to become loose.  This in turn caused the trim wheels to wobble  slightly generating undue noise.  

Solution:

Tighten the two nuts at the end of the rod that holds the two trim wheels in place.

(B) Flaps 5 Not Engaging

The problem with the flaps 5 micro-button has been discussed in an earlier post.  To summarize, when you moved the flaps lever to flaps 5 the correct flaps were not selected on the aircraft or registered on the PoKeys 55 interface card.  Several hours were spent checking connections, micro-buttons, wiring and the custom VGA cables that connect the flaps section of the quadrant to the appropiate interface module; the problem could not be discovered.  

Solution:

One of the two Belkin powered hubs located within the IMM had been replaced with another powered unit.  It appears the problem was that the replacement hub had too low a voltage, as a replacement with a higher voltage solved the problem.

(C) Throttle Thrust Levers Not Synchronizing (A/T on)

The two thrust levers of the quadrant did not synchronize when the Auto Throttle (A/T) was engaged; one lever would always be ahead or behind of the other.  At other times they would split apart (do the splits) when A/T was engaged.  

Solution:

The problem was easily solved by altering the tension on the slip clutch nut.  When the nut was  tightened, both levers moved together as one unit.  The secret was finding the appropriate torque.

(D) Throttle Thrust Levers Difficult To Move in Manual Mode (A/T Off)

The ability to move the thrust levers in manual mode (Autothrottle turned off) was not fluid and the levers occasionally snagged or were sticky when trying to move them.  

The fan belt is barely visible linking the pulley of the motor to the main pulley inside the quadrant

This is caused by the fan belt not moving smoothly through the groove of the pulley wheel.   The Autothrottle when engaged overrides any stickiness due to the power and torque of the Auto Throttle motor.

Solution:

Unfortunately, there isn’t a lot you can do to rectify this issue as it’s a by-product of using a mechanical system in which the fan belt is central to the consistent operation of the unit.  

The conundrum is that if you tighten the fan belt too much you will be unable to manually move the thrust levers as they will be exceptionally stiff and difficult to move (as you are pushing against the tension of the fan belt); however, if you loosen the fan belt too much, although the levers will move fluidly by hand, the fan belt may not have enough tension to move the levers when Auto throttle is engaged.  It’s a matter of compromise; selecting an appropriate in-between tension to allow acceptable manual and Autothrottle operation.

A more reliable method is to use a small gearbox, a simple slip clutch and a coupler to connect to the spur gear.  Another option is to use an electrical system.

Further thought needs to be done in this area before a decision is made to replace the fan belt system.  If a new system is incorporated, the change-out will be documented in a future post.

Conclusion

Despite some of the shortcomings to this conversion, in particular the mechanical fan belt system, the throttle unit shows a marked improvement on the earlier 300 series conversion. 

Since the project began there has been three throttle conversions, and wth each conversion has built upon knowledge learnt from earlier conversions.  Initially there was the 737-300 conversion in 2012, which was converted in a rudimentary way and only operated in manual mode.  This was followed by the conversion of the 737-500 throttle in 2016.  This throttle was then rebuilt and upgraded in 2017.

Further Information

  • A summary of the articles that address the conversion of the 737-500 series throttle quadrant conversion, and the rebuild and update can be found in Flight Controls/Throttle Quadrant.

Acronyms and Glossary

  • AFDS - Autopilot Flight Director system

  • A/T – Autothrottle

  • CMD A/B - Autopilot on/off for system A or system B

  • Flight Avionics Software - Sim Avionics, ProSim737 or similar

  • FMC - Flight Management Computer

  • MCP - Main Control Panel

  • QAMP – Quick Access Mounting Plate

  • Throttle Arc – The arc of the thrust levers from the end of the blocks to fully forward.  The term refers to the curved piece of aluminum that the throttle levers are moved along

  • TO/GA - Takeoff Go-around switch

  • %N1 -  Very simply explained, %N1 is throttle demand and as N1 (and N2) spin at absurdly high speeds, it is easier to simply reference a percentage and display that to the crew. It's much easier for our brains to interpret a value on a scale of 0-100% rather than tens of thousands of RPM